15:01:01 <bswartz> #startmeeting manila
15:01:02 <openstack> Meeting started Thu Sep 29 15:01:01 2016 UTC and is due to finish in 60 minutes.  The chair is bswartz. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:03 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:01:04 <bswartz> hello all
15:01:07 <openstack> The meeting name has been set to 'manila'
15:01:09 <xyang1> hi
15:01:11 <ganso> hello
15:01:13 <tbarron> hi
15:01:17 <markstur_> hi
15:01:18 <vponomaryov> hello
15:01:20 <jseiler> hi
15:01:25 <dustins> \o
15:01:34 <rraja> hi
15:02:10 <bswartz> so a few people are out it looks like
15:02:36 <toabctl> hi
15:02:40 <bswartz> #agenda https://wiki.openstack.org/wiki/Manila/Meetings
15:02:56 <bswartz> nothing on the agenda other than more ocata design summit planning
15:03:24 <bswartz> we're 1 week away from final release of newton, and I haven't heard about any release blocker bugs
15:03:51 <bswartz> has anyone seen any bugs that might be worthy of an rc2?
15:04:16 * bswartz hears silence
15:04:20 <bswartz> excellent
15:04:34 <tbarron> manila-ui is handled separately, right?
15:04:34 <bswartz> #topic Ocata Design Summit
15:04:37 <vponomaryov> manila is excellent ))
15:04:57 <tbarron> we have CVE fix to get into manila-ui for newton
15:05:03 <bswartz> tbarron: it's considered a "lib" and it will get a bugfix release
15:05:18 <tbarron> bswartz: k, thx
15:05:24 <bswartz> tbarron: just need to backport to stable/newton branch
15:05:43 <tbarron> so we need some love for https://review.openstack.org/#/c/378972/
15:05:52 <tbarron> and for https://review.openstack.org/#/c/378972/
15:06:04 <tbarron> we need the second for the first to pass
15:06:08 <bswartz> why is jenkins unhappy?
15:06:12 <vponomaryov> it is the same )
15:06:18 <tbarron> just told you :)
15:06:19 <ganso> tbarron: both links are the same
15:06:22 <bswartz> tbarron: same link
15:06:24 <tbarron> ah, sorry
15:06:33 <vponomaryov> #link https://review.openstack.org/378971
15:06:35 <tbarron> https://review.openstack.org/#/c/378971/
15:06:41 <tbarron> vponomaryov: ty
15:06:54 <bswartz> okay
15:07:11 <bswartz> #link https://etherpad.openstack.org/p/manila-ocata-design-summit-topics
15:07:45 <bswartz> so thanks for proposing some topics here
15:08:02 <bswartz> these topics seem to mostly overlap with what I was going to propose
15:08:48 <bswartz> however the thing that bothers me the most seems to be missing still
15:09:03 <bswartz> I would like to spend a bunch of time eliminating race conditions
15:09:48 <bswartz> I tossed out a proposal for how to do it a few month back, but I get the feeling it will require a concerted effort to actually make progress
15:09:49 <ganso> bswartz: if I am not mistaken there is a topic proposed by goutham in the list wrt to that
15:09:56 <bswartz> ganso: which?
15:10:01 <bswartz> the DLM one?
15:10:22 <tbarron> related, but not exactly the same
15:10:28 <ganso> bswartz: below "access rules fixes"
15:10:51 <bswartz> yeah there is an important distinction between DLM stuff and race condition elimination
15:11:03 <ganso> bswartz: "micro-states"
15:11:19 <bswartz> okay I see it in a sub-sub bullet
15:11:34 <bswartz> microstates and lock deserve top level topics
15:11:43 <bswartz> I want to spend at least a whole working session on this stuff, maybe more
15:11:49 <tbarron> +1
15:11:58 <bswartz> a DLM is worthless if we don't even have locks where we need them
15:12:01 * ganso is on his smartphone thus is typing slowly
15:12:34 <bswartz> so my proposal is to prioritize getting the locking right, and how we tackle distributed environments is a follow on concern
15:13:01 <bswartz> of course we need to make sure we don't pick a solution that precludes distributed/scalable environments
15:13:07 <ganso> bswartz: we should have a list of problems beforehand
15:13:12 <bswartz> however I think my proposal already meets that requirement
15:13:14 <bswartz> ganso: yes
15:13:21 <bswartz> that's where I was going
15:13:35 <bswartz> I want to write down several concrete problems we face today, and outline concrete solutions
15:14:03 <ganso> bswartz: we kinda discussed a bit of that in Austin, but we were unable to cite which problems exactly besides the access rules one which is going to be fixed by redesign
15:14:11 <xyang1> tbarron: I thought you have a WIP patch on DLM.  are you still working on it?
15:14:11 <bswartz> and I want to make sure everyone is on the same page about the solutions so we don't end up with disagreements about the right way to do state transitions
15:14:31 <tbarron> xyang1: not currently
15:14:33 <bswartz> ganso: that's fair
15:14:44 <bswartz> xyang1: you're thinking of gouthamr
15:15:03 <xyang1> bswartz: I thought tbarron proposed a patch
15:15:24 <bswartz> ganso: I think the access rules issue happened to be the one that gets exposed by the gate most often
15:15:28 <tbarron> xyang1: and I agree with bswartz that getting the model for handliing races is the first step, mechanism later
15:15:48 <xyang1> tbarron: sure
15:15:54 <bswartz> however we have other horrible race conditions -- such as create snapshot racing with delete share
15:16:31 <vponomaryov> bswartz: actually, I don't think such race is a problem
15:16:46 <vponomaryov> bswartz: there is no use case for it
15:17:08 <bswartz> vponomaryov: there is no use case for it, but it needs to fail predictably
15:17:25 <vponomaryov> bswartz: race of several access rules - it si a problem, yes
15:17:27 <toabctl> vponomaryov, use case? if one user creates a snapshot and another one deletes the share...
15:17:35 <xyang1> bswartz: this is tbarron's patch: https://review.openstack.org/#/c/318336
15:17:39 <bswartz> create snapshot can even race with itself, which is not good
15:18:02 <bswartz> xyang1: author/commiter are both gouthamr
15:18:05 <tbarron> xyang1: that's mostly goutham's at this point
15:18:10 <xyang1> ok
15:18:22 <vponomaryov> toabctl: yes, it is not use case, they shoot at each other
15:18:51 <toabctl> vponomaryov, and that shouldn't be possible imo.
15:18:56 <bswartz> I'm sure if we dig we will find other race conditions
15:19:13 <vponomaryov> toabctl: I would say such race is low priority
15:19:14 <bswartz> the point is that these are not hard to fix by proper state transitions
15:19:26 <vponomaryov> toabctl: very low
15:19:31 <bswartz> and we can give reasonable error messages when the bad races occur
15:20:23 <bswartz> anyways I'll add my topic proposal to the list
15:20:35 <bswartz> but I wanted to spend this time going over any other proposals
15:21:48 <bswartz> vponomaryov: I like how you want features to be implemented, but not by you :-)
15:22:27 <bswartz> do any of these topics need discussing?
15:22:28 <ganso> bswartz: we are not going to have a "remaining topics session" like on Austin friday, right?
15:22:41 <bswartz> ganso: that's a good question
15:23:03 <bswartz> unlike Austin there will be contributor meetups for only half the day
15:23:13 <vponomaryov> bswartz: I would just +1 if anyone proposed topics I mentioned that I do not own )
15:23:31 <bswartz> and we can use the contributor meetup time for whatever we like
15:23:57 <bswartz> we could use that time to hit a bunch of smaller topics, or we could use that time to dig in on one very hard topic
15:24:31 <bswartz> ganso: did you want a "remining topics" session?
15:25:07 <ganso> bswartz: not sure, it just seemed like it aacording to the schedule you linked last week
15:25:13 <tbarron> I like re-mining old topics
15:25:26 <bswartz> I think one mistake we made in Austin was that we discussed a lot of stuff which nobody ultimately spent any time on in Newton
15:25:53 <bswartz> since Ocata is going to be so squeezed for time, we need to make sure we use our time productively and talk about stuff which will actually get done
15:25:57 <ganso> bswartz: and we got to choose the most relevant ones for the sessions
15:26:20 <bswartz> ganso: yeah that's how we've done it before
15:26:37 <bswartz> ganso: I think that results in people choosing "interesting" topics though even if there's no willpower to make them happen
15:28:07 <bswartz> so perhaps we should make sure to capture commitments from people to actually get stuff done before we put a topic on the schedule
15:28:46 <bswartz> btw I looked at the ocata schedule and there only about 8 non-holiday weeks between design summit and feature freeze
15:29:19 <bswartz> so make sure you think about that when signing up to implement anything
15:29:34 <bswartz> of course there's a few weeks before the design summit where we can get work done
15:29:36 <markstur__> so the good news is we're almost done with Ocata
15:29:54 <markstur__> that's coming pretty fast
15:30:02 <bswartz> indeed
15:30:04 <ganso> markstur__: lol
15:30:49 <bswartz> markstur is being funny but it's not a joke -- we really need to think in terms of taking stuff that's almost done and finishing it up for ocata
15:30:56 <bswartz> starting something new is most likely doomed to fail
15:31:08 <bswartz> unless it's very small
15:32:41 <bswartz> okay so anything else on design summit?
15:32:58 <bswartz> I will add my topics to the etherpad when I'm back in the office next week
15:33:18 <bswartz> #topic open discussion
15:33:31 <bswartz> dustins: I saw that you sent out a ML post about docs
15:33:40 <bswartz> dustins: thanks for doing that
15:33:43 <ganso> what about https://review.openstack.org/#/c/374883/ ?
15:34:04 <bswartz> ganso: yeah I need to update that when I'm back in the office too
15:34:13 <ganso> bswartz: ok
15:34:45 <bswartz> ganso: I could do a round of typo-fixing real quick today so people can offer more substantive feedback
15:35:27 <tbarron> that's exactly what we need to avoid (the first, not the second)
15:35:33 <tbarron> fix all the typos at the end
15:35:46 <ganso> bswartz: I am not sure typos are holding back that many people
15:35:58 <tbarron> seeing 43 iterations on a spec with no changes but typos ...
15:36:02 <bswartz> tbarron: well all the comments create clutter which could discourage others from replying
15:36:25 <bswartz> tbarron: there's a reason I've waited this long
15:36:31 <tbarron> bswartz: true that, so do many iterations with no substantive changes
15:36:40 <ganso> bswartz: seems there are concerns to be addressed, not really so many wrt to typos
15:37:08 <bswartz> ganso: well those I won't get to today, unless we want to discuss them here
15:37:54 <ganso> bswartz: np, yes we could discuss, assuming everyone has read it by now. But I am fine with discussing on the spec
15:38:09 <ganso> s/spec/gerrit
15:38:20 <bswartz> okay so look for a new revision Monday
15:38:34 <bswartz> I'll take tbarron's advice and leave typos in
15:38:46 <bswartz> :-D
15:39:18 <bswartz> did dustins vanish?
15:39:20 <vponomaryov> bswartz: make typos in most valueable places ))
15:39:39 <tbarron> dustins is quite busy today
15:39:56 * dustins switches back to meeting
15:40:00 <bswartz> I wanted to know if he's getting any help with the docs items he asked for help on
15:40:00 <tbarron> :)
15:40:06 <dustins> bswartz: You're welcome :)
15:40:22 <dustins> There's still a LOT to do, but it looks like there's been some activity
15:40:36 <bswartz> okay well I'll offer a plug here
15:40:41 <dustins> So thanks to the folks that have been adding to the etherpad
15:40:45 <bswartz> #link https://etherpad.openstack.org/p/newton_documentation_blitz
15:41:16 <bswartz> thanks for those that are helping keep the docs in shape
15:41:22 <bswartz> they matter a lot to actual users
15:41:58 <dustins> +1000
15:42:35 <bswartz> alright so just a reminder that final release isn't for 1 more week
15:43:09 <bswartz> if a critical bug is found, we still have time to get a fix into the release, so keep hunting for bugs and make sure to call attention to any that seem worth of RC2 consideration
15:43:37 <bswartz> otherwise, the focus should be on updating docs and preparing for ocata
15:43:53 <bswartz> that's all for today
15:44:05 <bswartz> #endmeeting